Ultrasonic diagnostic device and storage medium

Abstract

To provide technology that can reduce a frequency of automatic changes of imaging conditions with timing unintended by a user.SOLUTION: An ultrasonic diagnostic device comprises: an ultrasonic probe; and a processor configured to communicate with the ultrasonic probe. The processor is configured to: set imaging conditions for acquiring ultrasonic images of a subject; transmit an ultrasonic beam to the ultrasonic probe and cause the ultrasonic probe to receive an echo from the subject in accordance with the imaging conditions, and then generate an ultrasonic image of the subject on the basis of the echo received by the ultrasonic probe; identify an imaging site included in the ultrasonic image; and determine whether to change the imaging conditions on the basis of the identified imaging site. Upon identifying that the ultrasonic image includes a plurality of imaging sites (thyroid, carotid), the processor does not change the imaging conditions.SELECTED DRAWING: Figure 11

Description

[Technical field]

[0001] The present invention relates to an ultrasonic diagnostic apparatus capable of changing imaging conditions, and a storage medium including instructions to be executed by the ultrasonic diagnostic apparatus. [Background technology]

[0002] When scanning a subject using an ultrasonic diagnostic apparatus, a user sets imaging conditions for each imaging region before starting a scan of the subject. [Prior art documents] [Patent documents]

[0003] [Patent Document 1] Patent Publication No. 2021-171559 Summary of the Invention [Problem to be solved by the invention]

[0004] The imaging conditions include various parameters. Therefore, it is difficult for the user to select the optimal parameters for each imaging region. Therefore, the ultrasound diagnostic device is provided with presets that define imaging conditions in advance for each imaging region. When imaging a subject, the user can set imaging conditions corresponding to the imaging region by selecting a preset corresponding to the imaging conditions of the subject, and thus can obtain high-quality ultrasound images.

[0005] However, depending on the user, there are cases where they are unable to select an appropriate preset or where the parameters are not adequately adjusted according to the part to be imaged, making it difficult to examine the subject under appropriate imaging conditions.

[0006] As a method for dealing with this problem, a technology has been considered that uses deep learning technology to determine the subject's imaging site based on the subject's ultrasound image, and automatically changes the imaging conditions if the current imaging conditions set by the user are not suitable for the subject's imaging site. However, depending on the subject's imaging site, it may be difficult to determine the site, and the subject's imaging site may not be correctly determined. Therefore, as the number of times the imaging conditions are automatically changed increases, there is a problem that the frequency with which the imaging conditions are automatically changed at times unintended by the user also increases.

[0007] Therefore, it is desirable to provide a technique that can reduce the frequency with which the shooting conditions are automatically changed without the user's intention. [Means for solving the problem]

[0008] A first aspect of the present invention is an ultrasonic probe, a processor in communication with the ultrasound probe; An ultrasound diagnostic device comprising: The processor, Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run The processor, When it is determined that the ultrasound image includes a plurality of imaging regions, the ultrasound diagnostic apparatus determines not to change the imaging conditions.

[0009] A second aspect of the present invention is an ultrasonic probe, a processor in communication with the ultrasound probe and a storage device; 1. An ultrasound system comprising: The processor, Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run The storage device stores a database including a plurality of categories, each of the plurality of categories including a plurality of parameter sets; The processor, identifying a first category corresponding to the identified imaging region from among the plurality of categories; identifying a second category corresponding to the set photographing condition from among the plurality of categories; determining not to change the photographing conditions when the first category and the second category match; This is an ultrasound diagnostic device that performs the following:

[0010] A third aspect of the present invention is a recording medium having stored thereon instructions executable by a processor in communication with an ultrasound probe, the instructions causing the processor to: Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run the command, The instructions further include for the processor to: The storage medium determines not to change the imaging conditions when it is determined that the ultrasound image includes a plurality of imaging regions.

[0011] A fourth aspect of the present invention is a recording medium having stored thereon instructions executable by a processor in communication with an ultrasound probe, the instructions causing the processor to: Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run the command, The determination of whether to change the imaging condition is identifying a first category corresponding to the identified imaging region from among the plurality of categories; identifying a second category corresponding to the set photographing condition from among the plurality of categories; determining not to change the photographing conditions when the first category and the second category match; A storage medium comprising: Effect of the Invention

[0012] In one aspect of the present invention, when it is determined that the ultrasound image includes multiple imaging regions, it is determined that the imaging conditions are not changed. Therefore, it is possible to reduce the frequency of the imaging conditions being automatically changed at a timing not intended by the user. It is also possible to reduce the risk that the currently set imaging conditions are changed to imaging conditions that are not suitable for the imaging regions.

[0013] In another aspect of the present invention, the parameter sets are classified by category, and when the first category corresponding to the identified imaging region and the second category corresponding to the set imaging condition match, it is determined that the imaging condition will not be changed. Therefore, it is possible to reduce the frequency of the imaging condition being automatically changed at a timing not intended by the user. It is also possible to reduce the risk that the currently set imaging condition will be changed to an imaging condition that is not suitable for the imaging region. [Brief description of the drawings]

[0014] [Figure 1] FIG. 1 is a diagram showing a state in which a subject is scanned by an ultrasound diagnostic apparatus 1 according to a first embodiment of the present invention. [Diagram 2] FIG. 1 is a block diagram of an ultrasonic diagnostic device 1. [Diagram 3] FIG. 2 is a schematic diagram of an original image. [Figure 4] FIG. [Diagram 5] FIG. 13 is an explanatory diagram of labeling of correct answer data. [Figure 6] FIG. 1 is a schematic explanatory diagram of a method for creating a trained model. [Figure 7] FIG. 1 is a diagram showing an example of a flowchart executed in an examination of a subject. [Figure 8] FIG. 13 is an explanatory diagram of inference of an imaging region. [Figure 9] FIG. 13 is an explanatory diagram of an example in which it is inferred that an input image 62 includes a plurality of imaging regions. [Figure 10] FIG. 2 is an explanatory diagram of a database. [Figure 11] FIG. 1 is a diagram showing an example of a flowchart executed in an examination of a subject. [Figure 12] FIG. 11 is an explanatory diagram of step ST33. [Figure 13] FIG. 11 is an explanatory diagram of step ST33. [Figure 14] FIG. 13 is an explanatory diagram of a training image used to create a trained model of the third embodiment. [Figure 15]FIG. 13 is an explanatory diagram of parameter sets according to breast size. [Figure 16] FIG. 1 is a diagram showing an example of a flowchart executed in an examination of a subject. [Figure 17] FIG. 11 is an explanatory diagram of step ST33. [Figure 18] FIG. 11 is an explanatory diagram of step ST33. [Figure 19] FIG. 13 is an explanatory diagram of a training image used to create a trained model in the fourth embodiment. [Figure 20] FIG. 11 is an explanatory diagram of a parameter set to be registered. [Figure 21] FIG. 1 is a diagram showing an example of a flowchart executed in an examination of a subject. [Figure 22] FIG. 11 is an explanatory diagram of step ST33. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0015] Hereinafter, a mode for carrying out the invention will be described, however, the present invention is not limited to the following mode.

[0016] (1) First embodiment FIG. 1 is a diagram showing a state in which a subject is scanned by an ultrasound diagnostic apparatus 1 according to a first embodiment of the present invention, and FIG.

[0017] The ultrasound diagnostic device 1 includes an ultrasound probe 2, a transmit beamformer 3, a transmitter 4, a receiver 5, a receive beamformer 6, a processor 7, a display unit 8, a memory 9, and a user interface 10.

[0018] The ultrasonic probe 2 has a plurality of transducer elements 2a arranged in an array. The transmit beamformer 3 and the transmitter 4 drive the plurality of transducer elements 2a arranged in the ultrasonic probe 2, and ultrasonic waves are transmitted from the transducer elements 2a. The ultrasonic waves transmitted from the transducer elements 2a are reflected inside the subject, and the reflected echoes are received by the transducer elements 2a. The transducer elements 2a convert the received echoes into electrical signals, and output these electrical signals as echo signals to the receiver 5. The receiver 5 performs a predetermined process on the echo signals, and outputs them to the receive beamformer 6. The receive beamformer 6 performs receive beamforming on the signals received from the receiver 5, and outputs echo data.

[0019] The receive beamformer 6 may be a hardware beamformer or a software beamformer. If the receive beamformer 6 is a software beamformer, the receive beamformer 6 may comprise one or more processors including one or more of: i) a graphics processing unit (GPU); ii) a microprocessor; iii) a central processing unit (CPU); iv) a digital signal processor (DSP); v) other types of processors capable of performing logical operations. The processors constituting the receive beamformer 6 may be separate from the processor 7 or may be comprised in the processor 7.

[0020] The ultrasound probe 2 may include electrical circuitry for performing all or part of the transmit beamforming and / or receive beamforming. For example, the transmit beamformer 3, the transmitter 4, the receiver 5, and all or part of the receive beamformer 6 may be provided within the ultrasound probe 2.

[0021] The processor 7 controls the transmit beamformer 3, the transmitter 4, the receiver 5, and the receive beamformer 6. The processor 7 is also in electronic communication with the ultrasound probe 2. The processor 7 controls which of the transducer elements 2a are active and the shape of the ultrasound beam transmitted from the ultrasound probe 2. The processor 7 is also in electronic communication with the display 8 and the user interface 10. The processor 7 can process the echo data to generate an ultrasound image. The term "electronic communication" can be defined to include both wired and wireless communication. The processor 7 can include a central processing unit (CPU) according to one embodiment. According to other embodiments, the processor 7 can include one or more processors or other electronic components capable of performing processing functions, such as a digital signal processor, a field programmable gate array (FPGA), a graphics processing unit (GPU), or other types of processors. According to other embodiments, the processor 7 can include multiple electronic components capable of performing processing functions. For example, the processor 7 can include two or more electronic components selected from a list of electronic components including a central processing unit, a digital signal processor, a field programmable gate array, and a graphics processing unit.

[0022] The processor 7 may also include a complex demodulator (not shown) for demodulating the RF data. In alternative embodiments, the demodulation may be performed earlier in the processing chain.

[0023] The processor 7 can also generate various ultrasound images (e.g., B-mode images, color Doppler images, M-mode images, color M-mode images, spectral Doppler images, elastography images, TVI images, strain images, strain velocity images, etc.) based on data obtained by processing by the receive beamformer 6. One or more modules can also generate these ultrasound images.

[0024] The image beams and / or image frames may be stored and timing information recorded indicating when the data was acquired in memory. The modules may include, for example, a scan conversion module that performs a scan conversion operation to convert the image frames from coordinate beam space to display space coordinates. A video processor module may be provided that reads the image frames from the memory and displays the image frames in real time while a procedure is being performed on the subject. The video processor module may store the image frames in an image memory, and the ultrasound images are read from the image memory and displayed on a display 8.

[0025] As used herein, the term "image" may refer broadly to both a visible image and data representing a visible image, and the term "data" may include raw data, which is ultrasound data before a scan conversion operation, and image data, which is data after a scan conversion operation.

[0026] The above-mentioned processing tasks handled by the processor 7 may be executed by a plurality of processors. Furthermore, when the receive beamformer 6 is a software beamformer, the processing executed by the beamformer may be executed by a single processor or may be executed by multiple processors.

[0027] The display unit 8 is, for example, an LED (Light Emitting Diode) display unit, an LCD (Liquid Crystal Display), or an organic EL (Electro-Luminescence) display unit. The display unit 8 displays an ultrasound image. As shown in FIG. 1, the display unit 8 includes a display monitor 18 and a touch panel 28, but the display unit 8 may be configured as a single display unit instead of the display monitor 18 and the touch panel 28. Also, instead of the display monitor 18 and the touch panel 28, two or more display devices may be provided.

[0028] The memory 9 is any known data storage medium. In one example, the ultrasound image display system includes a non-transient storage medium and a transient storage medium as the memory. The ultrasound image display system may also include a plurality of memories. The non-transient storage medium is, for example, a non-volatile storage medium such as a hard disk drive (HDD) or a read only memory (ROM). The non-transient storage medium may include a portable storage medium such as a compact disk (CD) or a digital versatile disk (DVD). The program executed by the processor 7 is stored in the non-transient storage medium. The transient storage medium is a volatile storage medium such as a random access memory (RAM).

[0029] The memory 9 stores one or more instructions executable by the processor 7. The one or more instructions cause the processor 7 to perform operations described in the first to fourth embodiments described below.

[0030] The processor 7 may be configured to be connected to an external storage device via a wired or wireless connection. In this case, the instructions to be executed by the processor 7 may be stored in both the memory 9 and the external storage device in a distributed manner.

[0031] The user interface 10 can accept input from the user 51. For example, the user interface 10 accepts input of instructions and information from the user 51. The user interface 10 includes a keyboard, hard keys, a trackball, a rotary control, soft keys, and the like. The user interface 10 may include a touch screen (for example, the touch screen of the touch panel 28) that displays the soft keys and the like. The ultrasonic diagnostic apparatus 1 is configured as described above.

[0032] When scanning a subject using an ultrasonic diagnostic apparatus, a user sets imaging conditions for each imaging region before starting a scan of the subject.

[0033] The imaging conditions include various parameters. Therefore, it is difficult for the user to select the optimal parameters for each imaging region. Therefore, the ultrasound diagnostic device is provided with presets that define imaging conditions in advance for each imaging region. When imaging a subject, the user can set imaging conditions corresponding to the imaging region by selecting a preset corresponding to the imaging conditions of the subject.

[0034] However, depending on the user, there are cases where they are unable to select an appropriate preset or where the parameters are not adequately adjusted according to the part to be imaged, making it difficult to examine the subject under appropriate imaging conditions.

[0035] As a method for dealing with this problem, a technology has been considered that uses deep learning technology to determine the subject's imaging site based on the subject's ultrasound image, and automatically changes the imaging conditions if the current imaging conditions set by the user are not suitable for the subject's imaging site. However, depending on the subject's imaging site, it may be difficult to determine the site, and the subject's imaging site may not be correctly determined. Therefore, as the number of times the imaging conditions are automatically changed increases, there is a problem that the frequency with which the imaging conditions are automatically changed at times unintended by the user also increases.

[0036] Therefore, the ultrasound diagnostic device 1 of the first embodiment is configured to reduce the frequency of automatic changes in the imaging conditions at a timing not intended by the user. The first embodiment will be specifically described below.

[0037] In the first embodiment, the imaging site of the subject is inferred using the trained model, and whether or not to change the imaging conditions is determined based on the inference result. Therefore, in the first embodiment, before the subject is examined, a learning phase for generating a trained model is executed to generate a trained model suitable for inferring the imaging site of the subject. Therefore, in the following, first, the learning phase for generating the trained model will be described. Then, after describing the learning phase, a method for automatically changing the imaging conditions during the examination of the subject will be described.

[0038] (About the learning phase) 3 to 6 are explanatory diagrams of the learning phase. In the learning phase, first, original images are prepared as the basis for generating training images.

[0039] FIG. 3 is a schematic diagram of the original image. In the first embodiment, a plurality of ultrasound images Ei (i=1 to n) are prepared as original images. The plurality of ultrasound images Ei (i=1 to n) are, for example, ultrasound images acquired at a medical facility such as a hospital or a medical equipment manufacturer. These ultrasound images Ei include images of various imaging regions. The imaging regions can be any region that can be the subject of ultrasound diagnosis, for example, "abdomen," "breast," "carotid artery," "thyroid gland," and "lower limbs," but are not limited to these regions. For example, 5,000 to 10,000 original images are prepared.

[0040] Next, pre-processing is performed on these original images Ei as shown in FIG. Preprocessing includes, for example, image cutting, standardization, normalization, image inversion, image rotation, magnification rate change, image quality change, etc. By preprocessing the original image Ei, a preprocessed original image EAi can be obtained. Each preprocessed original image is used as a training image for creating a trained model. Next, this training data is labeled with the correct answer data (see Figure 5).

[0041] FIG. 5 is a schematic explanatory diagram of labeling of the correct answer data. In the first embodiment, a plurality of imaging regions to be examined by a plurality of ultrasound diagnostic apparatuses 1 are used as the correct answer data.

[0042] For example, if the photographed part of the training image is the abdomen, the training image is labeled with "abdomen" as correct answer data, and if the photographed part of the training image is the breast, the training image is labeled with "breast" as correct answer data.

[0043] Next, a trained model is created using the above training data. Figure 6 is a schematic diagram illustrating a method for creating a trained model.

[0044] The trained model 31 can be created by having the neural network 30 learn the above-mentioned training images. The trained model 31 is stored in a memory or an external storage device. The trained model 31 can be created using any learning algorithm used in AI learning, machine learning, deep learning, or the like. For example, the trained model 31 may be created by supervised learning or unsupervised learning.

[0045] In the first embodiment, it is determined whether or not to automatically change the shooting conditions by using the trained model 31. An example of this determination method will be described below with reference to FIG.

[0046] FIG. 7 is a diagram showing an example of a flowchart executed in an examination of a subject. In step ST1, a user 51 (e.g., a doctor or an ultrasound technician) guides a subject to an examination room and has the subject lie on an examination bed. The user 51 also operates a user interface 10 (see FIG. 2) to input patient information, set imaging conditions for acquiring an ultrasound image of the subject, and perform other necessary settings. The imaging conditions include any conditions related to acquiring an ultrasound image, such as transmission conditions of an ultrasound beam, reception conditions of echoes from the subject, and data processing conditions used to create an ultrasound image based on the received echoes.

[0047] In this example, the imaging site of the subject is assumed to be the "thyroid gland." Therefore, the user 51 sets imaging conditions for the thyroid gland.

[0048] When the user 51 has completed preparation for the examination, the user 51 starts the examination of the subject. In Fig. 7, the examination start time is indicated as t0.

[0049] 7, "subject", "imaged part", and "imaging conditions" are shown on the time axis. "Subject" indicates the subject being examined, "imaged part" indicates the imaged part of the subject, and "imaging conditions" indicate the imaging conditions set in the ultrasound diagnostic device. For example, at the start time t0 of the examination, the "subject" is the subject S11, the "imaged part" is the thyroid gland, and the "imaging conditions" are imaging conditions V1 for the thyroid gland.

[0050] The user 51 operates the ultrasound probe 2 while pressing it against the imaging region of the subject S11 to scan the subject S11. In this case, since the imaging region of the subject S11 is the thyroid gland, the user 51 performs the examination by pressing the ultrasound probe 2 against the neck of the subject S11. The ultrasound probe 2 transmits ultrasound waves and receives echoes reflected within the subject S11. The received echoes are converted into electrical signals, and the electrical signals are output as echo signals to the receiver 5 (see FIG. 2). The receiver 5 performs a predetermined process on the echo signals and outputs them to the receive beamformer 6. The receive beamformer 6 performs receive beamforming on the signals received from the receiver 5 and outputs echo data. The processor 7 generates an ultrasound image based on the echo data.

[0051] The user 51 checks the generated ultrasound image and, if necessary, stores the ultrasound image, and then continues to perform the examination of the subject S11.

[0052] On the other hand, after the examination of the subject S11 is started at time t0, the processor 7 periodically executes a process 41 for determining whether to change the imaging conditions and automatically changing the imaging conditions as necessary. In this embodiment, the first process 41 is executed at time t1 after the examination start time t0. This process 41 will be described below.

[0053] When the process 41 starts, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound image acquired between time points t0 and t1. This determination step ST10 will be described below. First, in step ST11, the processor generates an input image 62 to be input to the trained model 31 based on an ultrasound image 61 acquired between time t0 and time t1. Specifically, the processor 7 generates the input image 62 by preprocessing the ultrasound image 61. This preprocessing is basically the same as the preprocessing executed when generating the training image of the trained model 31. By executing the preprocessing, the input image 62 to be input to the trained model 31 can be generated.

[0054] When one ultrasound image 61 is acquired between time t0 and time t1, the processor 7 generates an input image 62 to be input to the trained model 31 based on the ultrasound image 61.

[0055] On the other hand, when multiple ultrasound images are acquired between time t0 and time t1, the processor 7 selects one ultrasound image 61 from the multiple ultrasound images, and generates an input image 62 for inputting to the trained model 31 based on the selected ultrasound image 61. When multiple ultrasound images are acquired between time t0 and time t1, the processor 7 can typically select as the ultrasound image 61 the last acquired ultrasound image or the ultrasound image acquired immediately before time t1 at which the process 41 is started. After generating the input image 62, the process proceeds to step ST12.

[0056] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 62 (see FIG. 8).

[0057] FIG. 8 is an explanatory diagram of inference of an imaging region. The processor 7 inputs the input image 62 to the trained model 31, and infers the parts included in the input image 62 using the trained model 31. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 62. Then, the processor 7 infers the imaging part included in the input image 62 based on the probability calculated for each imaging part.

[0058] The processor 7 compares the probability calculated for each imaging part with a threshold value. The threshold value is a reference value for determining whether each imaging part is included in the input image 62. In this embodiment, if the probability of an imaging part is equal to or greater than the threshold value, the imaging part is inferred to be included in the input image 62. For example, if only the thyroid gland among multiple imaging parts has a probability exceeding the threshold value, the processor 7 infers that the imaging part included in the input image 62 is the thyroid gland, as shown in FIG. 8. On the other hand, there are also cases where it is inferred that multiple imaging parts are included in the input image 62 (see FIG. 9).

[0059] FIG. 9 is an explanatory diagram of an example in which it is inferred that the input image 62 includes a plurality of imaging regions. For example, if the input image 62 includes both the thyroid gland and the carotid artery, and the probabilities of the thyroid gland and the carotid artery exceed a threshold, the processor 7 infers that the input image 62 includes two imaging regions, namely the thyroid gland and the carotid artery.

[0060] Here, it is assumed that the trained model 31 outputs "thyroid gland" as an inference result, as shown in Fig. 8. Therefore, in step ST12, the processor 7 infers that the imaging part is the thyroid gland. After inferring the imaging part, the process proceeds to step ST20.

[0061] In step ST20, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST20 will be specifically described below.

[0062] First, in step ST21, the processor 7 judges whether the number of imaging parts inferred for the input image 62 is one or two or more. If the number of imaging parts is one, the process proceeds to step ST22, and if the number of imaging parts is two or more, the process proceeds to step ST23. Here, since it is inferred that the input image 62 includes only the thyroid gland, the processor 7 determines that the number of imaging parts included in the input image 62 is one. Therefore, the process proceeds to step ST22.

[0063] In step ST22, the processor 7 determines whether or not to change the imaging conditions. The processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. If the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12, the processor 7 decides not to change the imaging conditions and ends the process 41. On the other hand, if the currently set imaging conditions are not imaging conditions corresponding to the imaging region inferred in step ST12, the process proceeds to step ST24, where the processor 7 decides to change the imaging conditions.

[0064] At time t1, the set imaging condition is the imaging condition V1 for the thyroid gland. Meanwhile, the imaging part inferred in step ST12 is the thyroid gland. Therefore, since the currently set imaging condition V1 is the imaging condition corresponding to the imaging part (thyroid gland) inferred in step ST12, the processor 7 decides not to change the imaging condition and ends the process 41.

[0065] Meanwhile, the user 51 continues to examine the subject S11 by operating the ultrasound probe 2 even after time t1. The processor 7 periodically executes the above process 41 during the examination of the subject S11. Here, it is assumed that the imaging of the thyroid gland of the subject S11 is completed without changing the imaging conditions. The end time of the imaging of the thyroid gland of the subject S11 is indicated as "t2". When the examination of the subject S11 is completed, preparations are made for the examination of the next subject S12.

[0066] In the following, a case will be described in which the imaging site of the new subject S12 is different from that of the previous subject S11. Here, the imaging site of the previous subject S11 was the thyroid gland, but the imaging site of the new subject S12 is the carotid artery.

[0067] After the user 51 has finished the thyroid examination of the previous subject S11, the user 51 prepares for the examination of the carotid artery of a new subject S12. In this case, the imaging part is changed from the thyroid to the carotid artery, so the user 51 needs to change the imaging condition from the imaging condition V1 for the thyroid to the imaging condition V2 for the carotid artery. However, in the following, a case will be considered where the user 51 starts the examination of the carotid artery of the new subject S12 without changing the imaging condition.

[0068] At time t3, the user 51 starts an examination of the carotid artery of a new subject S12. The user 51 starts the examination of the carotid artery of the subject S12 from time t3, but since the imaging conditions have not been changed, the set imaging conditions remain the thyroid imaging conditions V1. Therefore, the user 51 starts the examination of the carotid artery of the subject S12 under the thyroid imaging conditions V1.

[0069] On the other hand, after the examination of the carotid artery of the subject S12 is started at time t3, the processor 7 periodically executes the process 41. In this embodiment, a case where the process 41 is executed at time t4 after time t3 will be described.

[0070] When the process 41 starts, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t3 and t4. This determination step ST10 will be described below.

[0071] First, in step ST11, the processor 7 pre-processes the ultrasound image 63 acquired between time t3 and time t4 to generate an input image 64 to be input to the trained model 31.

[0072] When one ultrasound image 63 is acquired between time t3 and time t4, the processor 7 can preprocess the ultrasound image 63 to generate an input image 64 to be input to the trained model 31. On the other hand, when multiple ultrasound images are acquired between time t3 and time t4, the processor 7 can select one ultrasound image 63 from the multiple ultrasound images and generate an input image 64 to be input to the trained model 31 based on the selected ultrasound image 63. When multiple ultrasound images are acquired between time t3 and time t4, the processor 7 can typically select the ultrasound image last acquired between time t3 and time t4 (the ultrasound image acquired immediately before time t4) as the ultrasound image 63. After generating the input image 64, the process proceeds to step ST12.

[0073] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 64. The processor 7 inputs the input image 64 to the trained model 31, and infers the part included in the input image 64 using the trained model 31. Here, it is assumed that the processor 7 infers that the input image 64 includes the carotid artery, as shown in Fig. 7. After inferring the part to be imaged, the process proceeds to step ST20.

[0074] In step ST20, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST20 will be specifically described below.

[0075] First, in step ST21, the processor 7 judges whether the number of imaging parts inferred for the input image 64 is one or two or more. If the number of imaging parts is one, the process proceeds to step ST22, and if the number of imaging parts is two or more, the process proceeds to step ST23. Here, since it is inferred that the input image 64 includes only the carotid artery, the processor 7 determines that the number of imaging parts included in the input image 64 is one. Therefore, the process proceeds to step ST22.

[0076] In step ST22, the processor 7 determines whether or not to change the imaging conditions. The processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. If the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12, the processor 7 decides not to change the imaging conditions and ends the process 41. On the other hand, if the currently set imaging conditions are not imaging conditions corresponding to the imaging region inferred in step ST12, the process proceeds to step ST24, where the processor 7 decides to change the imaging conditions.

[0077] At time t4, the set imaging condition is the imaging condition V1 for the thyroid gland. On the other hand, the imaging part inferred in step ST12 is the carotid artery. Therefore, the currently set imaging condition is not the imaging condition corresponding to the imaging part (carotid artery) inferred in step ST12. Therefore, the processor 7 decides to change the imaging condition, proceeds to step ST24, and changes the imaging condition from the imaging condition V1 for the thyroid gland to the imaging condition V2 for the carotid artery. Therefore, immediately after the flow of process 41 is completed (immediately after time t4), the imaging condition is automatically changed from the imaging condition V1 for the thyroid gland to the imaging condition V2 for the carotid artery.

[0078] Therefore, the user 51 starts imaging the carotid artery of the subject S12 without changing the imaging conditions to the imaging conditions V2 for the carotid artery, but the processor 7 sets the imaging conditions to the imaging conditions V2 for the carotid artery immediately after the time point t4. Therefore, even if the user 51 forgets to change the imaging conditions, after the processor 7 changes the imaging conditions, the user 51 can image the carotid artery of the subject S12 according to the imaging conditions V2 for the carotid artery, and therefore a high-quality image of the carotid artery can be acquired.

[0079] After time t4, the user 51 continues to inspect the carotid artery of the subject S12 while operating the ultrasound probe 2, and the processor 7 periodically executes the above process 41. An example in which the processor 7 executes the process 41 again at time t5 will be described.

[0080] When the process 41 starts, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t4 and t5. This determination step ST10 will be described below. First, in step ST11, the processor pre-processes an ultrasound image 65 acquired between time t4 and time t5 to generate an input image 66 to be input to the trained model 31.

[0081] When one ultrasound image 65 is acquired between time t4 and time t5, the processor 7 can preprocess the ultrasound image 65 to generate an input image 66 to be input to the trained model 31. On the other hand, when multiple ultrasound images are acquired between time t4 and time t5, the processor 7 can select one ultrasound image 65 from the multiple ultrasound images and generate an input image 66 to be input to the trained model 31 based on the selected ultrasound image 65. When multiple ultrasound images are acquired between time t4 and time t5, the processor 7 can typically select the ultrasound image last acquired between time t4 and time t5 (the ultrasound image acquired immediately before time t5) as the ultrasound image 65. After generating the input image 66, the process proceeds to step ST12.

[0082] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 66. The processor 7 inputs the input image 66 to the trained model 31, and infers the part included in the input image 66 using the trained model 31. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 66. Then, the processor 7 infers the imaging part included in the input image 66 based on the probability calculated for each imaging part.

[0083] Here, it is assumed that the probability of the thyroid gland and the probability of the carotid artery exceed the threshold. Therefore, in this case, the processor 7 infers that the input image 66 includes the thyroid gland and the carotid artery, as shown in Fig. 7. After inferring the imaging site, the process proceeds to step ST20.

[0084] In step ST20, first, step ST21 is executed. In step ST21, the processor 7 judges whether the number of imaging sites inferred for the input image 66 is one or two or more. If the number of imaging sites is one, the process proceeds to step ST22, and if the number of imaging sites is two or more, the process proceeds to step ST23. Here, since it is inferred that the input image 66 includes the thyroid gland and the carotid artery, the processor 7 determines that the input image 66 includes two or more imaging sites. If it is determined that the input image 66 includes two or more imaging sites, the process proceeds to step ST23, where the processor 7 determines not to change the imaging conditions. The reason for this will be described below.

[0085] It is inferred that the input image 66 includes two imaging sites, that is, the thyroid gland and the carotid artery. Therefore, when changing the imaging conditions, the processor 7 must first identify whether the imaging site of the subject S12 is the thyroid gland or the carotid artery. However, at time t5, the user 51 is examining the carotid artery. If the processor 7 identifies that the current imaging site is the thyroid gland and automatically changes the imaging conditions to imaging conditions for the thyroid gland, the user 51 will continue to image the carotid artery of the subject S12 under the imaging conditions for the thyroid gland. Therefore, if the processor 7 does not change the imaging conditions, the user 51 could have continued imaging the carotid artery of the subject S12 under the imaging conditions for the carotid artery. However, because the processor 7 has changed the imaging conditions, the user 51 will continue to image the carotid artery of the subject S12 under the imaging conditions for the thyroid gland.

[0086] In this embodiment, in order to prevent such a problem from occurring, when the processor 7 infers that the input image 66 includes a plurality of imaging regions, the processor 7 determines not to change the imaging conditions in step ST23 and ends the flow of the process 41. Therefore, when the processor 7 infers that the input image 66 includes a plurality of imaging regions, the processor 7 does not change the imaging conditions, so that it is possible to avoid a risk that the currently set imaging conditions are automatically changed to imaging conditions for another imaging region that is not intended by the user 51.

[0087] On the other hand, even after time t5, the user 51 continues to examine the carotid artery of the subject S12 while operating the ultrasound probe 2, and the processor 7 periodically executes the above process 41. Then, when all images required for diagnosing the carotid artery of the subject S12 are acquired, the examination ends (time t7).

[0088] In this embodiment, when the input image includes multiple imaging regions, the imaging conditions are not changed and the process 41 is terminated. Therefore, it is possible to reduce the frequency of the imaging conditions being automatically changed at a timing not intended by the user 51. It is also possible to avoid the risk of the current imaging conditions being automatically changed to imaging conditions for another imaging region not intended by the user 51.

[0089] (2) Second embodiment In the second embodiment, an example will be described in which a parameter set included in the imaging conditions of an imaging region is classified into a plurality of categories, and whether or not to change the imaging conditions is determined based on the classification result.

[0090] In the second embodiment, a database for managing a plurality of categories is stored in the memory (or an external storage device) of the ultrasound diagnostic device. In describing the second embodiment, the database for managing a plurality of categories will be described first. After this description, a flow for automatically changing the imaging conditions will be described.

[0091] FIG. 10 is an explanatory diagram of a database stored in the memory (or an external storage device) of the ultrasound diagnostic device. A plurality of categories 1 to N are registered in the database. Each category includes a plurality of parameter sets. Each parameter set includes a plurality of parameters that are set when imaging a region to be imaged. Each parameter set is classified into one of the plurality of categories 1 to N based on whether the anatomical features of the region to be imaged are similar and / or whether the parameter values ​​are similar. Categories 1 to N in the database will be described below by taking categories 1 and 2 out of categories 1 to N. An explanatory diagram of the parameter sets of categories 1 and 2 is shown in the right half of FIG. 10.

[0092] Category 1 includes a thyroid parameter set A1 and a carotid parameter set A2. The thyroid parameter set A1 includes a plurality of parameters that are set during thyroid imaging. In FIG. 10, the parameters included in the thyroid parameter set A1 are (a 11 , a 12 , a 13 , a 1z These parameters are shown as a 11 ~a 1z For example, the frequency and the depth are included.

[0093] The carotid artery parameter set A2 includes a plurality of parameters that are set when imaging the carotid artery. In FIG. 10, the parameters included in the carotid artery parameter set A2 are (a 21 , a 22 , a 23 , a 2z These parameters are shown as a 21 ~a 2z For example, the frequency and the depth are included.

[0094] Category 2 includes an upper arm vein parameter set B1, an upper arm artery parameter set B2, a lower limb vein parameter set B3, and a lower limb artery parameter set B4.

[0095] The upper arm vein parameter set B1 includes a plurality of parameters that are set when imaging the upper arm vein. In FIG. 10, the parameters included in the upper arm vein parameter set B1 are (b 11 , b 12 , b 13 , b 1z These parameters are shown in b 11 ~b 1z For example, the frequency and the depth are included.

[0096] The brachial artery parameter set B2 includes a plurality of parameters that are set when imaging the brachial artery. In FIG. 10, the parameters included in the brachial artery parameter set B2 are (b 21 , b 22 , b 23 , b 2zThese parameters are shown in b 21 ~b 2z For example, the frequency and the depth are included.

[0097] The lower limb vein parameter set B3 includes a plurality of parameters that are set when imaging the lower limb veins. In FIG. 10, the parameters included in the lower limb vein parameter set B3 are (b 31 , b 32 , b 33 , b 3z These parameters are shown in b 31 ~b 3z For example, the frequency and the depth are included.

[0098] The lower limb artery parameter set B4 includes a plurality of parameters that are set when imaging the lower limb arteries. In FIG. 10, the parameters included in the lower limb artery parameter set B4 are (b 41 , b 42 , b 43 , b 4z These parameters are shown in b 41 ~b 4z For example, the frequency and the depth are included.

[0099] Although a detailed description is omitted, the other categories 3 to N also include parameter sets each including a plurality of parameters.

[0100] The above parameters may be preset before an object is examined, or may be manually set by the user 51 before an object is examined.

[0101] In the second embodiment, an example will be described in which it is determined whether or not to change the imaging conditions during an examination of a subject by utilizing the above categories 1 to N registered in a database.

[0102] FIG. 11 is a diagram showing an example of a flowchart executed in an examination of a subject. In step ST1, the user 51 guides the subject to an examination room and has the subject lie on an examination bed. The user 51 also operates the user interface 10 (see FIG. 2) to input patient information, set imaging conditions for acquiring an ultrasound image of the subject, and perform other necessary settings. Here, it is assumed that the subject's imaging site is the "brachial artery." Therefore, the user 51 sets imaging conditions for the brachial artery.

[0103] When the user 51 has completed preparation for the examination, the user 51 starts the examination of the subject. In Fig. 11, the examination start time is indicated as t0.

[0104] 11, "subject", "imaged part", and "imaging conditions" are shown on the time axis. "Subject" indicates the subject being examined, "imaged part" indicates the imaged part of the subject, and "imaging conditions" indicate the imaging conditions set in the ultrasound diagnostic device. For example, at the start time t0 of the examination, the "subject" is subject S21, the "imaged part" is the brachial artery, and the "imaging conditions" are imaging conditions V1 for the brachial artery.

[0105] A user 51 operates the ultrasonic probe 2 while pressing it against the upper arm of the subject S21 to examine the subject S21. The ultrasonic probe 2 transmits ultrasonic waves and receives echoes reflected within the subject S21. The received echoes are converted into electrical signals, and these electrical signals are output as echo signals to a receiver 5 (see FIG. 2). The receiver 5 performs a predetermined process on the echo signals and outputs them to a receive beamformer 6. The receive beamformer 6 performs receive beamforming on the signals received from the receiver 5 and outputs echo data. The processor 7 generates an ultrasound image based on the echo data.

[0106] The user 51 checks the generated ultrasound image and, if necessary, stores the ultrasound image, and then continues to perform the examination of the subject.

[0107] On the other hand, after the examination of the subject S21 is started at time t0, the processor 7 periodically executes a process 42 for determining whether to change the imaging conditions and automatically changing the imaging conditions as necessary. In this embodiment, the first process 42 is executed at time t1 after the examination start time t0. This process 42 will be described below.

[0108] When the process 42 is started, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t0 and t1. This determination step ST10 will be described below.

[0109] First, in step ST11, the processor 7 generates an input image 72 to be input to the trained model 31 by preprocessing an ultrasound image 71 acquired between time t0 and time t1.

[0110] When one ultrasound image 71 is acquired between time t0 and time t1, the processor 7 generates an input image 72 to be input to the trained model 31 based on the ultrasound image 71. The processor 7 generates the input image 72, for example, by preprocessing the ultrasound image 71 acquired immediately before time t1. After generating the input image 72, the process proceeds to step ST12.

[0111] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 72. The processor 7 inputs the input image 72 to the trained model 31, and infers the part included in the input image 72 using the trained model 31. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 72. Then, the processor 7 infers the imaging part included in the input image 72 based on the probability calculated for each imaging part.

[0112] Here, it is assumed that the probability of the brachial artery exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 72 is the brachial artery. After inferring the imaging region, the process proceeds to step ST30.

[0113] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0114] First, in step ST31, the processor 7 judges whether or not the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. If the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12, the process proceeds to step ST32, whereas if the currently set imaging conditions are not imaging conditions corresponding to the imaging region inferred in step ST12, the process proceeds to step ST33.

[0115] At time t1, the set imaging conditions are those for the brachial artery. Meanwhile, the imaging region inferred in step ST12 is the brachial artery. Therefore, the currently set imaging conditions are those corresponding to the imaging region (brachial artery) inferred in step ST12. Therefore, the process proceeds to step ST32, where the processor 7 decides not to change the imaging conditions, and ends the process 42.

[0116] Meanwhile, the user 51 continues to examine the subject S21 while operating the ultrasound probe 2 even after time t1. The processor 7 periodically executes the above process 42 during the examination of the subject S21. Here, it is assumed that the imaging of the brachial artery of the subject S21 is completed without changing the imaging conditions. The end time of the examination of the brachial artery of the subject S21 is indicated as "t2". When the examination of the subject S21 is completed, preparations are made for the examination of the next new subject S22.

[0117] In the following, a case will be described in which the imaging site of the new subject S22 is different from that of the previous subject S21. Here, the imaging site of the previous subject S21 was the brachial artery, but the imaging site of the new subject S22 is the vein of the lower limb.

[0118] After completing the examination of the brachial artery of the previous subject S21, the user 51 prepares to examine the lower limb veins of a new subject S22. In this case, the imaging site is changed from the brachial artery to the lower limb veins, so the user 51 needs to change the imaging conditions from the imaging conditions for the brachial artery to the imaging conditions for the lower limb veins. However, in the following, a case will be considered where the user 51 starts the examination of the lower limb veins of the new subject S22 without changing the imaging conditions.

[0119] The user 51 starts an examination of the veins of the lower extremities of a new subject S22. Here, the start time of the examination of the veins of the lower extremities of the new subject S22 is indicated as t3.

[0120] The user 51 starts an examination of the veins of the lower extremities of the subject S22 from time t3, but since the imaging conditions have not been changed, the set imaging conditions remain the imaging conditions W1 for the brachial artery. Therefore, the user 51 starts an examination of the veins of the lower extremities of the new subject S22 under the imaging conditions W1 for the brachial artery.

[0121] On the other hand, after the examination of the lower limb veins of a new subject S22 is started at time t3, the processor 7 periodically executes the process 42. In this embodiment, a case where the process 42 is executed at time t4 after time t3 will be described.

[0122] First, in step ST11, the processor 7 preprocesses the ultrasound image 73 acquired between time t3 and time t4 to generate an input image 74 to be input to the trained model 31. The processor 7 generates the input image 74 by, for example, preprocessing the ultrasound image 73 acquired immediately before time t4. After generating the input image 74, the process proceeds to step ST12.

[0123] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 74. The processor 7 inputs the input image 74 to the trained model 31, and infers the part included in the input image 74 by using the trained model 31. Here, it is assumed that the processor 7 infers that the input image 74 includes veins of the lower limbs, as shown in Fig. 11. After inferring the part to be photographed, the process proceeds to step ST30.

[0124] In step ST30, the processor 7 determines whether or not to change the conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0125] First, in step ST31, the processor 7 determines whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t4, the set imaging conditions are imaging conditions for the brachial artery. On the other hand, the imaging region inferred in step ST12 is the lower limb artery. Therefore, at time t4, the set imaging conditions (imaging conditions for the brachial artery) are not imaging conditions corresponding to the imaging region inferred in step ST12 (lower limb vein), and therefore the process proceeds to step ST33.

[0126] FIG. 12 is an explanatory diagram of step ST33. In step ST33, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. Here, since the inferred imaging region is "lower limb vein", the processor 7 identifies category 2 including a parameter set for the lower limb vein from among the plurality of categories 1 to N.

[0127] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for the brachial artery, the processor 7 identifies category 2 including a parameter set for the brachial artery from among a plurality of categories 1 to N. After identifying category 2, the process proceeds to step ST34.

[0128] In step ST34, the processor 7 determines whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories match (category 2), so the process proceeds to step ST32.

[0129] In step ST32, the processor 7 decides not to change the imaging conditions. That is, at time t4, the lower limb veins of the subject S22 are imaged under the imaging conditions for the brachial artery, not the imaging conditions for the lower limb veins, but if the categories specified in step ST33 match, the processor 7 decides not to change the imaging conditions. The reason for this will be described below.

[0130] Category 2 includes an upper arm vein parameter set B1, an upper arm artery parameter set B2, a lower limb vein parameter set B3, and a lower limb artery parameter set B4. The upper arm vein, the upper arm artery, the lower limb vein, and the lower limb artery, which are the imaging sites of these parameter sets, are anatomically similar to each other. In addition, the parameter values ​​used in imaging the upper arm vein, the upper arm artery, the lower limb vein, and the lower limb artery are often the same or close to each other. Therefore, it is considered that the image quality of the ultrasound image will not be significantly affected even if an ultrasound image of the lower limb vein is acquired under the imaging conditions for the upper arm artery without changing the currently set imaging conditions for the upper arm artery to the imaging conditions for the inferred lower limb vein. Therefore, if the categories identified in step ST33 match, it is determined that the imaging conditions will not be changed. A plurality of parameter sets included in one category are parameter sets for a plurality of sites defined as having similar body compositions of the subject. The body composition includes fat, muscle, bone, and blood vessels. Even if the parameter set is changed between a plurality of parameter sets included in one category, an ultrasound image of mutually acceptable image quality can be obtained.

[0131] Therefore, the user 51 continues to inspect the lower extremity veins of the subject S22 under the brachial artery imaging conditions after time t4. Meanwhile, the processor 7 periodically executes the process 42 after time t4. Here, it is assumed that the process 42 is executed after time t4, but it is determined not to change the imaging conditions (step ST32). Therefore, the inspection of the lower extremity veins of the subject S22 ends without automatic change of the imaging conditions. In FIG. 11, the end time of the inspection of the lower extremity veins of the subject S22 is indicated as t5. When the inspection of the subject S22 ends, the next new subject S23 starts to be inspected. It is assumed that the imaging site of the new subject S23 is the thyroid gland.

[0132] After finishing the previous examination of the subject S22, the user 51 prepares for the examination of the thyroid gland of the subject S23. In this case, the imaging part is changed from the lower limb vein to the thyroid gland, so the user 51 needs to change the imaging conditions from the imaging conditions for the brachial artery to the imaging conditions for the thyroid gland. However, in the following, a case will be considered where the user 51 starts the examination of the thyroid gland of a new subject S23 without changing the imaging conditions.

[0133] The user 51 starts a thyroid examination of a new subject S23. Here, the start time of the thyroid examination of the new subject S23 is indicated as t6.

[0134] The user 51 starts an examination of the thyroid gland of a new subject S23 at time t6, but does not change the imaging conditions, so the set imaging conditions remain the imaging conditions for the brachial artery. Therefore, the user 51 starts an examination of the thyroid gland of the subject S23 under the imaging conditions for the brachial artery.

[0135] On the other hand, after the thyroid examination of the subject S23 is started at time t6, the processor 7 periodically executes the process 42. In this embodiment, a case will be described in which the process 42 is executed at time t7 after time t6.

[0136] When the process 42 is started, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t6 and t7. This determination step ST10 will be described below.

[0137] First, in step ST11, the processor 7 preprocesses an ultrasound image 75 acquired between time t6 and time t7 to generate an input image 76 to be input to the trained model 31. The processor 7 generates the input image 76, for example, by preprocessing an ultrasound image 75 acquired immediately before time t7. After generating the input image 76, the process proceeds to step ST12.

[0138] In step ST12, the processor 7 uses the trained model 31 to infer the part represented by the input image 76. The processor 7 inputs the input image 76 to the trained model 31, and infers the part included in the input image 76 using the trained model 31. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 76. Then, the processor 7 infers the imaging part included in the input image 76 based on the probability calculated for each imaging part.

[0139] Here, it is assumed that the probability of the thyroid gland exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 76 is the thyroid gland. After inferring the imaging region, the process proceeds to step ST30.

[0140] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0141] First, in step ST31, the processor 7 determines whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t7, the set imaging conditions are imaging conditions for the brachial artery. On the other hand, the imaging region inferred in step ST12 is the thyroid gland. Therefore, at time t7, the set imaging conditions (imaging conditions for the brachial artery) are not imaging conditions corresponding to the imaging region (thyroid gland) inferred in step ST12, so the process proceeds to step ST33.

[0142] FIG. 13 is an explanatory diagram of step ST33. In step ST33, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. In this example, since the inferred imaging region is "thyroid gland," the processor 7 identifies category 1 including a parameter set for the thyroid gland from among the plurality of categories 1 to n.

[0143] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for the brachial artery, the processor 7 identifies category 2 including a parameter set for the brachial artery from among a plurality of categories 1 to N. After identifying category 2, the process proceeds to step ST34.

[0144] In step ST34, the processor 7 determines whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories do not match (categories 1 and 2), so the process proceeds to step ST35.

[0145] In step ST35, the processor 7 determines to change the shooting conditions. That is, when it is determined that the categories do not match, the processor 7 determines to change the shooting conditions. The reason for this will be described below.

[0146] Categories 1 and 2 are classified by anatomically dissimilar imaging sites. Therefore, imaging sites belonging to category C1 and imaging sites belonging to category C2 are imaging sites that are easy to distinguish from each other because of their significant anatomical differences. In addition, when comparing imaging sites in category 1 and imaging sites in category 2, the parameter values ​​set in the imaging conditions often have a large difference. Therefore, if the imaging conditions for the brachial artery set at time t7 are changed to the inferred imaging conditions for the thyroid gland and an ultrasound image of the thyroid gland is acquired, it is considered that the image quality of the acquired ultrasound image can be greatly improved. Therefore, if the two categories identified in step ST33 do not match, it is decided to change the imaging conditions.

[0147] When it is determined to change the imaging conditions, the process proceeds to step ST36, where the processor 7 changes the imaging conditions for the brachial artery set at time t7 to imaging conditions for the thyroid gland. For example, when changing the imaging conditions, the processor 7 can change the brachial artery parameter set B1 to the thyroid gland parameter set A1. In addition to changing the parameter set, the processor 7 may change other conditions that are not included in the parameter set.

[0148] Therefore, the user 51 can perform a thyroid examination of the subject S23 under the thyroid imaging conditions immediately after time t7. Meanwhile, even after time t7, the processor 7 periodically executes the above process 42. Then, when acquisition of all images required for the examination of the subject S23 is completed, the examination of the subject S23 is terminated (time t8).

[0149] In the second embodiment, as shown in FIG. 10, a plurality of categories 1 to N are registered in the database. Each category includes a plurality of parameter sets. Each parameter set is classified into one of the plurality of categories 1 to N based on whether the anatomical features of the imaging site are similar and / or whether the parameter values ​​are similar. For example, the upper arm vein, the upper brachial artery, the lower limb vein, and the lower limb artery are all blood vessels, so they are anatomically similar, and the parameter values ​​used in the imaging conditions are the same or similar. Therefore, for the upper arm vein, the upper brachial artery, the lower limb vein, and the lower limb artery, an ultrasound image of a certain quality can be obtained using any of the upper arm vein parameter set B1, the upper brachial artery parameter set B2, the lower limb vein parameter set B3, and the lower limb artery parameter set B4. Therefore, in the second embodiment, the upper arm vein parameter set B1, the upper brachial artery parameter set B2, the lower limb vein parameter set B3, and the lower limb artery parameter set B4 are classified into the same category 2. Therefore, since the imaging conditions are not changed between the upper arm vein, the upper arm artery, the lower leg vein, and the lower leg artery, it is possible to reduce the frequency of the imaging conditions being automatically changed at a timing unintended by the user 51. Furthermore, since the parameter sets B1 to B4 have the same or similar parameter values, it is also possible to acquire an ultrasound image of sufficient quality even if the imaging conditions are not changed between the upper arm vein, the upper arm artery, the lower leg vein, and the lower leg artery.

[0150] Moreover, the thyroid parameter set A1 and the carotid artery parameter set A2 are classified into the same category 1. Therefore, the imaging conditions are not changed between the thyroid gland and the carotid artery, so that it is possible to reduce the frequency of the imaging conditions being automatically changed at a timing unintended by the user 51. Furthermore, since the parameter sets A1 and A2 have the same or similar parameter values, it is possible to acquire an ultrasound image of sufficient quality even if the imaging conditions are not changed between the thyroid gland and the carotid artery.

[0151] Furthermore, in this embodiment, if the category specified in step ST33 is different, the imaging conditions are changed. Here, an example is shown in which the imaging conditions W1 for the brachial artery are changed to the imaging conditions W2 for the thyroid immediately after time point t7. Therefore, if the imaging regions are not anatomically similar or if the difference in parameter values ​​is large, the currently set imaging conditions can be automatically changed to imaging conditions suitable for the imaging region, so that a high-quality ultrasound image can be acquired.

[0152] (3) Third embodiment In the third embodiment, an example of imaging a breast will be described. In the third embodiment, an example of performing an ultrasound examination under imaging conditions according to the size of the breast will be described so that a higher quality breast image can be obtained.

[0153] In the third embodiment, an ultrasound examination is performed under imaging conditions according to the size of the breast, so a trained model capable of identifying the size of the breast is prepared. This trained model is created as follows.

[0154] FIG. 14 is an explanatory diagram of a training image used to create a trained model of the third embodiment.

[0155] First, training images of different breast sizes are prepared. Here, three types of training images of different breast sizes are prepared. The first are training images Q11-Q1z of small-sized breasts (hereinafter referred to as "breasts (S)"), the second are training images Q21-Q2z of standard-sized breasts (hereinafter referred to as "breasts (M)"), and the third are training images Q31-Q3z of large-sized breasts (hereinafter referred to as "breasts (L)").

[0156] Additionally, each training image is labeled with breast (S), breast (M), and breast (L) as ground truth data.

[0157] Then, the training images Q11-Q1z, Q21-Q2z, and Q31-Q3z are included in the training images used to create the trained model, and the neural network is trained to create a trained model that can distinguish between different breast sizes. Next, a parameter set according to the breast size is registered in a database.

[0158] FIG. 15 is a diagram illustrating parameter sets according to breast size. Category 3 includes a breast (S) parameter set C1, a breast (M) parameter set C2, and a breast (L) parameter set C3. These parameter sets C1 to C3 are registered in the database so as to be included in category 3.

[0159] The breast (S) parameter set C1 includes a plurality of parameters used when examining a subject having a small breast. In FIG. 15, the parameters included in the breast (S) parameter set C1 are (c 11 , c 12 , c 13 , c 1z These parameters are shown in c 11 ~c 1z For example, the frequency and the depth are included.

[0160] The breast (M) parameter set C2 includes a plurality of parameters used when examining a subject having a standard breast size. In FIG. 15, the parameters included in the breast (M) parameter set C2 are (c 21 , c 22 , c 23 , c 2z These parameters are shown in c 21 ~c 2z For example, the frequency and the depth are included.

[0161] The breast (L) parameter set C3 includes a plurality of parameters used when examining a subject having a large breast. In FIG. 15, the parameters included in the breast (L) parameter set C3 are (c 31 , c 32 , c 33 , c 3z These parameters are shown in c 31 ~c 3z For example, frequency and depth are included. Parameter sets C1, C2, and C3 are parameter sets set according to different compositions in one part (here, the breast). "Different compositions" include cases where the ratio of each element constituting the body composition and the combination of each element are different. Here, parameter sets C1, C2, and C3 are set according to the amount of fat in the breast, with the parameter sets C1, C2, and C3 having the largest amount of fat in that order. Even if the parameter set is changed between parameter sets C1, C2, and C3, it is possible to obtain ultrasound images of mutually acceptable image quality.

[0162] In the third embodiment, a test subject is inspected using the trained model 32 and the database. The test subject inspection flow will be described below.

[0163] FIG. 16 is a diagram showing an example of a flowchart executed in an examination of a subject. In step ST1, the user 51 guides the subject to an examination room and has the subject lie on an examination bed. The user 51 also operates the user interface 10 (see FIG. 2) to input patient information, set imaging conditions for acquiring an ultrasound image of the subject, and perform other necessary settings. Here, it is assumed that the imaging site of the subject is a breast (S) (a small-sized breast). Therefore, the user 51 sets imaging conditions for the breast (S).

[0164] When the user 51 has completed preparation for the examination, the user 51 starts the examination of the subject. In Fig. 16, the examination start time is indicated as t0.

[0165] 16, "subject", "imaged part", and "imaging conditions" are shown on the time axis. "Subject" indicates the subject being examined, "imaged part" indicates the imaged part of the subject, and "imaging conditions" indicate the imaging conditions set in the ultrasound diagnostic device. For example, at the start time t0 of the examination, the "subject" is the subject S31, the "imaged part" is the breast (S), and the "imaging conditions" are imaging conditions X1 for the breast (S).

[0166] A user 51 operates the ultrasound probe 2 while pressing it against the breast of the subject S31 to examine the subject S31. The ultrasound probe 2 transmits ultrasound waves and receives echoes reflected within the subject S31. The received echoes are converted into electrical signals, which are output as echo signals to a receiver 5 (see FIG. 2). The receiver 5 performs a predetermined process on the echo signals and outputs them to a receive beamformer 6. The receive beamformer 6 performs receive beamforming on the signals received from the receiver 5 and outputs echo data. The processor 7 generates an ultrasound image based on the echo data.

[0167] The user 51 checks the generated ultrasound image and, if necessary, stores the ultrasound image, and then continues to perform the examination of the subject. On the other hand, after the examination of the subject S31 is started at time t0, the processor 7 periodically executes a process 43 for determining whether to change the imaging conditions and automatically changing the imaging conditions as necessary. In this embodiment, the first process 43 is executed at time t1 after the examination start time t0. This process 43 will be described below.

[0168] When the process 43 starts, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t0 and t1. This determination step ST10 will be described below.

[0169] First, in step ST11, the processor 7 preprocesses an ultrasound image 81 acquired between time t0 and time t1 to generate an input image 82 to be input to the trained model 32. The processor 7 generates the input image 82, for example, by preprocessing an ultrasound image 81 acquired immediately before time t1. After generating the input image 82, the process proceeds to step ST12.

[0170] In step ST12, the processor 7 uses the trained model 32 to infer the part represented by the input image 82. The processor 7 inputs the input image 82 to the trained model 32, and infers the part included in the input image 82 using the trained model 32. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 82. Then, the processor 7 infers the imaging part included in the input image 82 based on the probability calculated for each imaging part.

[0171] Here, it is assumed that the probability of the breast (S) exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 82 is the breast (S). After inferring the imaging region, the process proceeds to step ST30.

[0172] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0173] First, in step ST31, the processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t1, the set imaging conditions are imaging conditions for the breast (S). On the other hand, the imaging region inferred in step ST12 is the breast (S). Therefore, the currently set imaging conditions are imaging conditions corresponding to the imaging region (breast (S)) inferred in step ST12. Therefore, the process proceeds to step ST32, where the processor 7 decides not to change the imaging conditions, and ends the process 43.

[0174] Meanwhile, the user 51 continues to examine the subject S31 while operating the ultrasound probe 2 even after time t1. The processor 7 periodically executes the above process 43 even after time t1. Here, it is assumed that imaging of the breast (S) of the subject S31 is completed without changing the imaging conditions. The end time of the examination of the breast (S) of the subject S31 is indicated as "t2".

[0175] When the examination of the subject S31 is completed, preparations are made for the examination of the next new subject S32. Here, the examination target of the new subject S32 is the breast, but the breast is assumed to be of standard size. In this case, the imaging region is changed from a small-sized breast ("breast (S)") to a standard-sized breast ("breast (M)"), so the user 51 needs to change the imaging conditions from those for the breast (S) to those for the breast (M). However, in the following, a case will be considered in which the user 51 starts the examination of the breast (M) of the new subject S32 without changing the imaging conditions.

[0176] The user 51 starts an examination of the breast (M) of a new subject S32. Here, the start time of the examination of the breast (M) of the new subject S32 is indicated as t3.

[0177] The user 51 starts an examination of the breast (M) of the subject S32 from time t3, but since the imaging conditions have not been changed, the set imaging conditions remain the imaging conditions for the breast (S). Therefore, the user 51 starts an examination of the breast (M) of the subject S32 under the imaging conditions for the breast (S).

[0178] On the other hand, after the examination of the breast (M) of the subject S32 is started at time t3, the processor 7 periodically executes the process 43. In this embodiment, a case where the process 43 is executed at time t4 after time t3 will be described.

[0179] First, in step ST11, the processor 7 preprocesses the ultrasound image 83 acquired between time t3 and time t4 to generate an input image 84 to be input to the trained model 32. The processor 7 generates the input image 84, for example, by preprocessing the ultrasound image 83 acquired immediately before time t4. After generating the input image 84, the process proceeds to step ST12.

[0180] In step ST12, the processor 7 infers a part represented by the input image 84 using the trained model 32. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 84. Then, the processor 7 infers the imaging part included in the input image 84 based on the probability calculated for each imaging part.

[0181] Here, it is assumed that the probability of a breast (M) (a normal-sized breast) exceeds the threshold. Therefore, the processor 7 infers that the imaging region included in the input image 84 is a breast (M). After inferring the imaging region, the process proceeds to step ST30.

[0182] In step ST30, the processor 7 determines whether or not to change the conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0183] First, in step ST31, the processor 7 determines whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t4, the set imaging conditions are imaging conditions for the breast (S). On the other hand, the imaging region inferred in step ST12 is the breast (M). Therefore, at time t4, the set imaging conditions (imaging conditions for the breast (S)) are not imaging conditions corresponding to the imaging region (breast (M)) inferred in step ST12, so the process proceeds to step ST33.

[0184] FIG. 17 is an explanatory diagram of step ST33. In step ST33, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. In this example, since the inferred imaging region is "breast (M)", the processor 7 identifies category 3 including a parameter set for breast (M) from among the plurality of categories 1 to n.

[0185] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for a breast (S), the processor 7 identifies category 3 including a parameter set for a breast (S) from among a plurality of categories 1 to N. After identifying category 3, the process proceeds to step ST34.

[0186] In step ST34, the processor 7 determines whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories match (category 3), so the process proceeds to step ST32.

[0187] In step ST32, the processor 7 decides not to change the imaging conditions. That is, at time t4, the breast (M) of the subject S32 is imaged under imaging conditions for the breast (S) rather than the imaging conditions for the breast (M), but if the category specified in step ST33 matches, it is decided not to change the imaging conditions. The reason for this will be described below.

[0188] As described above, the parameter sets C1, C2, and C3 are parameter sets set according to the size of the breast, but the organ to be imaged is the same for all of them, namely, "breast." Therefore, even if the breast (M) is imaged using the parameter set for the breast (S) or the parameter set for the breast (L), it is possible to obtain an ultrasound image of sufficient image quality. Therefore, in this embodiment, it is decided not to change the imaging conditions.

[0189] Therefore, the user 51 continues to examine the breast (M) of the subject S32 under the imaging conditions for the breast (S) even after time t4. Meanwhile, during the examination of the subject S32, the processor 7 periodically executes the above process 43. Here, it is assumed that the process 43 is executed after time t4 as well, but it is determined not to change the imaging conditions (step ST32). Therefore, the examination of the breast (M) of the subject S32 ends without automatically changing the imaging conditions. In FIG. 16, the end time of the examination of the breast (M) of the subject S32 is indicated as t5. When the examination of the subject S32 ends, the examination of the next new subject S33 begins. It is assumed that the imaging site of the new subject S33 is the thyroid gland.

[0190] After finishing the examination of the previous subject S32, the user 51 prepares for the examination of the thyroid gland of a new subject S33. In this case, the imaging part is changed from the breast to the thyroid gland, so the user 51 needs to change the imaging conditions from the imaging conditions for the breast (S) to the imaging conditions for the thyroid gland. However, in the following, a case will be considered where the user 51 starts the examination of the thyroid gland of the new subject S33 without changing the imaging conditions.

[0191] The user 51 starts a thyroid examination of a new subject S33. Here, the start time of the thyroid examination of the new subject S33 is indicated as t6.

[0192] The user 51 starts a thyroid examination of a new subject S33 at time t6, but does not change the imaging conditions, so the set imaging conditions remain the imaging conditions X1 for the breast (S). Therefore, the user 51 starts the thyroid examination of the subject S33 under the imaging conditions X1 for the breast (S).

[0193] On the other hand, after the thyroid examination of the subject S33 is started at time t6, the processor 7 periodically executes the process 43. In this embodiment, a case will be described in which the process 43 is executed at time t7 after time t6.

[0194] When the process 43 starts, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t6 and t7. This determination step ST10 will be described below.

[0195] First, in step ST11, the processor 7 preprocesses an ultrasound image 85 acquired between time t6 and time t7 to generate an input image 86 to be input to the trained model 32. The processor 7 preprocesses, for example, an ultrasound image 85 acquired immediately before time t7 to generate the input image 86. After generating the input image 86, the process proceeds to step ST12.

[0196] In step ST12, the processor 7 uses the trained model 32 to infer the part represented by the input image 86. The processor 7 inputs the input image 86 to the trained model 32, and infers the part included in the input image 86 using the trained model 32. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 86. Then, the processor 7 infers the imaging part included in the input image 86 based on the probability calculated for each imaging part.

[0197] Here, it is assumed that the probability of the thyroid gland exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 86 is the thyroid gland. After inferring the imaging region, the process proceeds to step ST30.

[0198] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0199] First, in step ST31, the processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t7, the set imaging conditions are imaging conditions for the breast (S). On the other hand, the imaging region inferred in step ST12 is the thyroid gland. Therefore, at time t7, the set imaging conditions (imaging conditions for the breast (S)) are not imaging conditions corresponding to the imaging region (thyroid gland) inferred in step ST12, so the process proceeds to step ST33.

[0200] FIG. 18 is an explanatory diagram of step ST33. In step ST33, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. Here, since the inferred imaging region is "thyroid gland", the processor 7 identifies category 1 including a parameter set for the thyroid gland from among the plurality of categories 1 to N.

[0201] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for a breast (S), the processor 7 identifies category 3 including a parameter set for a breast (S) from among a plurality of categories 1 to N. After identifying category 3, the process proceeds to step ST34.

[0202] In step ST34, the processor 7 determines whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories do not match (categories 1 and 3), so the process proceeds to step ST35.

[0203] In step ST35, the processor 7 determines to change the shooting conditions. That is, when it is determined that the categories do not match, the processor 7 determines to change the shooting conditions. The reason for this will be described below.

[0204] Categories 1 and 3 are classified according to anatomically dissimilar imaging sites. Therefore, imaging sites belonging to category 1 and imaging sites belonging to category 3 are imaging sites that are easy to distinguish from each other because of their significant anatomical differences. Furthermore, when comparing imaging sites in category 1 and those in category 3, there are often large differences in the parameter values ​​set in the imaging conditions. Therefore, it is considered that the image quality of the acquired ultrasound image can be greatly improved if the currently set imaging conditions of the breast (S) are changed to the inferred imaging conditions of the thyroid gland. Therefore, if the two categories identified in step ST33 do not match, it is decided to change the imaging conditions.

[0205] When it is determined to change the imaging condition, the process proceeds to step ST36, where the processor 7 changes the imaging condition for the breast (S) set at time t7 to the imaging condition for the thyroid. For example, when changing the imaging condition, the processor 7 can change the parameter set C1 for the breast (S) pulse to the parameter set A1 for the thyroid. In addition to changing the parameter set, the processor 7 may change other conditions that are not included in the parameter set.

[0206] Therefore, the user 51 can perform a thyroid examination of the subject S33 under the thyroid imaging condition X2 immediately after time point t7. Meanwhile, even after time point t7, the processor 7 periodically executes the above process 43. Then, when acquisition of all images required for the examination of the subject S33 is completed, the examination of the subject S33 is terminated (time point t8).

[0207] In the third embodiment, category 3 of the database includes a breast (S) parameter set C1, a breast (M) parameter set C2, and a breast (L) parameter set C3. The breast (S), breast (M), and breast (L) are common organs, and furthermore, the values ​​of parameters used in the imaging conditions are the same or similar. Therefore, for the breast (S), breast (M), and breast (L), an ultrasound image of a certain quality can be obtained using any of the parameter sets C1, C2, and C3. Therefore, in the third embodiment, the breast (S) parameter set C1, the breast (M) parameter set C2, and the breast (L) parameter set C3 are classified into the same category 3. By classifying the parameter sets C1, C2, and C3 into the same category 3, it is possible to prevent the imaging conditions from being changed between the breast (S), breast (M), and breast (L), and therefore it is possible to reduce the frequency of the imaging conditions being automatically changed at a timing unintended by the user 51. In addition, since the breast (S) parameter set C1, breast (M) parameter set C2, and breast (L) parameter set C3 have the same or similar parameter values, ultrasound images of sufficient quality can be obtained even if the shooting conditions are not changed between the breast (S), breast (M), and breast (L). (4) Fourth embodiment In the fourth embodiment, an example will be described in which imaging conditions for the liver, kidneys, and intestines are classified into the same category. For the liver, an example will be described in which imaging conditions for a healthy liver and imaging conditions for a lesioned liver are set separately in order to enable more useful diagnosis for each patient. Therefore, in the fourth embodiment, a trained model capable of distinguishing between a healthy liver and a lesioned liver is prepared. A method for creating this trained model will be described below.

[0208] FIG. 19 is an explanatory diagram of a training image used to create a trained model of the fourth embodiment. First, training images R11 to R1z of a healthy liver (hereinafter referred to as "liver (N)") and training images R21 to R2z of a diseased liver (hereinafter referred to as "liver (L)") are prepared as training images.

[0209] Furthermore, as training images, training images R31 to R3z of the kidneys and training images R41 to R4z of the intestines are prepared. Then, liver (N), liver (L), kidney, and intestine are labeled as ground truth data in each training image.

[0210] The training images used to create the trained model are then trained on the neural network using training images R11-R1z, R21-R2z, R31-R3z, and R41-R4z, creating a trained model that identifies healthy and diseased livers, kidneys, and intestines.

[0211] Next, the healthy liver parameter set, the diseased liver parameter set, the kidney parameter set, and the intestine parameter set are registered in a database.

[0212] FIG. 20 is an explanatory diagram of the parameter sets to be registered. Category 4 includes a liver (N) parameter set D1, a liver (L) parameter set D2, a kidney parameter set D3, and an intestine parameter set D4. These parameter sets D1 to D4 are registered in the database so as to be included in category 4.

[0213] The liver (N) parameter set D1 includes a plurality of parameters used when examining a healthy liver. In FIG. 20, the parameters included in the liver (N) parameter set D1 are (d 11 , d 12 , d 13 , d 1z These parameters are shown in 11 ~d 1zFor example, the frequency and the depth are included.

[0214] The liver (L) parameter set D2 includes a plurality of parameters used when examining the liver including a lesion. In FIG. 20, the parameters included in the liver (L) parameter set D2 are (d 21 , d 22 , d 23 , d 2z These parameters are shown in 21 ~d 2z For example, the frequency and the depth are included.

[0215] The kidney parameter set D3 includes a plurality of parameters used in performing kidney examinations. In FIG. 20, the parameters included in the kidney parameter set D3 are (d 31 , d 32 , d 33 , d 3z These parameters are shown in 31 ~d 3z For example, the frequency and the depth are included.

[0216] The intestine parameter set D4 includes a plurality of parameters used when examining the intestine. In FIG. 20, the parameters included in the intestine parameter set D4 are (d 41 , d 42 , d 43 , d 4z These parameters are shown in 41 ~d 4z For example, the frequency and the depth are included.

[0217] In the fourth embodiment, the above-described trained model and database are used to inspect a subject. The inspection flow of the subject will be described below.

[0218] FIG. 21 is a diagram showing an example of a flowchart executed in an examination of a subject. In step ST1, the user 51 guides the subject to an examination room and has the subject lie on an examination bed. The user 51 also operates the user interface 10 (see FIG. 2) to input patient information, set imaging conditions for acquiring an ultrasound image of the subject S33, and perform other necessary settings. Here, it is assumed that the imaging site of the subject is the liver (N). Therefore, the user 51 sets imaging conditions for the liver (N).

[0219] When the user 51 has completed preparation for the examination, the user 51 starts the examination of the subject. In Fig. 21, the examination start time point is indicated as t0.

[0220] In addition, in Fig. 21, "subject", "imaged part", and "imaging conditions" are shown on the time axis. "Subject" indicates the subject being examined, "imaged part" indicates the imaged part of the subject, and "imaging conditions" indicate the imaging conditions set in the ultrasound diagnostic device. For example, at the start time t0 of the examination, the "subject" is the subject S41, the "imaged part" is the liver (N), and the "imaging conditions" are imaging conditions Y1 for the liver (N).

[0221] A user 51 operates the ultrasound probe 2 while pressing it against the liver of the subject S41 to examine the subject S41. The ultrasound probe 2 transmits ultrasound waves and receives echoes reflected within the subject S41. The received echoes are converted into electrical signals, which are output as echo signals to a receiver 5 (see FIG. 2). The receiver 5 performs a predetermined process on the echo signals and outputs them to a receive beamformer 6. The receive beamformer 6 performs receive beamforming on the signals received from the receiver 5 and outputs echo data. The processor 7 generates an ultrasound image based on the echo data.

[0222] The user 51 checks the generated ultrasound image and, if necessary, stores the ultrasound image, and then continues to perform the examination of the subject.

[0223] On the other hand, after the examination of the subject S41 is started at time t0, the processor 7 periodically executes a process 44 for determining whether to change the imaging conditions and automatically changing the imaging conditions as necessary. In this embodiment, the first process 44 is executed at time t1 after the examination start time t0. This process 44 will be described below.

[0224] When the process 44 is started, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound image acquired between time points t0 and t1. This determination step ST10 will be described below.

[0225] First, in step ST11, the processor 7 preprocesses an ultrasound image 91 acquired between time t0 and time t1 to generate an input image 82 to be input to the trained model 33. The processor 7 preprocesses, for example, an ultrasound image 91 acquired immediately before time t1 to generate an input image 92. After generating the input image 92, the process proceeds to step ST12.

[0226] In step ST12, the processor 7 uses the trained model 33 to infer the part represented by the input image 92. The processor 7 inputs the input image 92 to the trained model 33, and infers the parts included in the input image 92 using the trained model 33. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 92. Then, the processor 7 infers the imaging part included in the input image 92 based on the probability calculated for each imaging part.

[0227] Here, it is assumed that the probability of the liver (N) exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 92 is the liver (N). After inferring the imaging region, the process proceeds to step ST30.

[0228] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0229] First, in step ST31, the processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t1, the set imaging conditions are imaging conditions for the liver (N). On the other hand, the imaging region inferred in step ST12 is the liver (N). Therefore, the currently set imaging conditions are imaging conditions corresponding to the imaging region (liver (N)) inferred in step ST12. Therefore, the process proceeds to step ST32, where the processor 7 decides not to change the imaging conditions, and ends the process 44.

[0230] Meanwhile, the user 51 continues to examine the subject S41 while operating the ultrasound probe 2 even after time t1. The processor 7 periodically executes the above process 44 even after time t1. Here, it is assumed that the imaging of the liver (N) of the subject S41 is completed without changing the imaging conditions. The end time of the examination of the liver (N) of the subject S41 is indicated as "t2".

[0231] When the examination of the subject S41 is completed, preparations are made for the examination of the next new subject S42. It is assumed here that the examination target of the new subject S42 is a liver with a lesion. In this case, the imaging site is changed from a healthy liver ("liver (N)") to a liver with a lesion ("liver (L)"), so the user 51 needs to change the imaging conditions from the imaging conditions Y1 for the liver (N) to the imaging conditions for the liver (L). However, in the following, a case will be considered where the user 51 starts the examination of the liver (L) of the new subject S42 without changing the imaging conditions.

[0232] The user 51 starts an examination of the liver (L) of a new subject S42. Here, the start time of the examination of the liver (L) of the new subject S42 is indicated as t3.

[0233] The user 51 starts an examination of the liver (L) of the subject S42 from time t3, but since the imaging conditions have not been changed, the set imaging conditions remain the imaging conditions for the liver (N). Therefore, the user 51 starts an examination of the liver (L) of the subject S42 under the imaging conditions for the liver (N).

[0234] On the other hand, after the examination of the liver (L) of the subject S42 is started at time t3, the processor 7 periodically executes the process 44. In this embodiment, a case where the process 44 is executed at time t4 after time t3 will be described.

[0235] First, in step ST11, the processor 7 preprocesses the ultrasound image 93 acquired between time t3 and time t4 to generate an input image 94 to be input to the trained model 33. The processor 7 generates the input image 94 by, for example, preprocessing the ultrasound image 93 acquired immediately before time t4. After generating the input image 94, the process proceeds to step ST12.

[0236] In step ST12, the processor 7 infers a part represented by the input image 94 using the trained model 33. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 94. Then, the processor 7 infers the imaging part included in the input image 94 based on the probability calculated for each imaging part.

[0237] Here, it is assumed that the probability of the liver (L) (liver with a lesion) exceeds the threshold. Therefore, the processor 7 infers that the imaging site included in the input image 94 is the liver (L). After inferring the imaging site, the process proceeds to step ST30.

[0238] In step ST30, the processor 7 determines whether or not to change the conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0239] First, in step ST31, the processor 7 determines whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t4, the set imaging conditions are imaging conditions for the liver (N). On the other hand, the imaging region inferred in step ST12 is the liver (L). Therefore, at time t4, the set imaging conditions (imaging conditions for the liver (N)) are not imaging conditions corresponding to the imaging region (liver (L)) inferred in step ST12, so the process proceeds to step ST33.

[0240] FIG. 22 is an explanatory diagram of step ST33. In step ST33, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. In this example, since the inferred imaging region is "liver (L)", the processor 7 identifies category 4 including a parameter set for the liver (L) from among the plurality of categories 1 to n.

[0241] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for the liver (N), the processor 7 identifies category 4 including the liver (N) parameter set from among a plurality of categories 1 to n. After identifying category 4, the process proceeds to step ST34.

[0242] In step ST34, the processor 7 judges whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories match (category 4), so the process proceeds to step ST32.

[0243] In step ST32, the processor 7 decides not to change the imaging conditions. That is, at time t4, the liver (L) of the subject S42 is imaged under imaging conditions for the liver (N) rather than the imaging conditions for the liver (L), but if the categories specified in step ST33 match, it is decided not to change the imaging conditions. The reason for this will be described below.

[0244] As described above, the liver (N) and the liver (L) have in common the organ to be imaged, that is, the liver. Therefore, even if the liver (L) is imaged using the parameter set of the liver (N), it is possible to obtain an ultrasound image of sufficient image quality. Therefore, in this embodiment, it is determined not to change the imaging conditions.

[0245] Therefore, the user 51 continues the examination of the liver (L) of the subject S42 under the imaging conditions for the liver (N) even after time t4. Meanwhile, the processor 7 periodically executes the above process 44 even after time t4. Here, it is assumed that the process 44 is executed after time t4, but it is determined not to change the imaging conditions (step ST32). Therefore, the examination of the liver (L) of the subject S42 ends without automatically changing the imaging conditions. In FIG. 21, the end time of the examination of the liver (L) of the subject S42 is indicated as t5. When the examination of the subject S42 ends, the examination of the next new subject S43 starts. It is assumed that the imaging site of the new subject S43 is the kidney.

[0246] After finishing the examination of the previous subject S42, the user 51 prepares for the examination of the kidney of the new subject S43. In this case, the imaging part is changed from the liver to the kidney, so the user 51 needs to change the imaging condition from the imaging condition Y1 for the liver (N) to the imaging condition for the kidney. However, in the following, a case will be considered where the user 51 starts the examination of the kidney of the new subject S43 without changing the imaging condition.

[0247] The user 51 starts a renal examination of a new subject S43. Here, the start time of the renal examination of the new subject S43 is indicated as t6.

[0248] The user 51 starts an examination of the kidneys of a new subject S43 at time t6, but does not change the imaging conditions, so the set imaging conditions remain the imaging conditions X1 for the liver (N). Therefore, the user 51 starts an examination of the kidneys of the subject S43 under the imaging conditions Y1 for the liver (N).

[0249] On the other hand, after the renal examination of the subject S43 is started at time t6, the processor 7 periodically executes the process 44. In this embodiment, a case where the process 44 is executed at time t7 after time t6 will be described.

[0250] When the process 44 is started, first, in step ST10, the processor 7 determines the imaging site included in the ultrasound images acquired between time points t6 and t7. This determination step ST10 will be described below.

[0251] First, in step ST11, the processor 7 preprocesses an ultrasound image 95 acquired between time t6 and time t7 to generate an input image 96 to be input to the trained model 33. The processor 7 can generate the input image 96, for example, by preprocessing an ultrasound image 95 acquired immediately before time t7. After generating the input image 96, the process proceeds to step ST12.

[0252] In step ST12, the processor 7 uses the trained model 33 to infer the part represented by the input image 96. The processor 7 inputs the input image 96 to the trained model 33, and infers the part included in the input image 96 using the trained model 33. In the inference step, the processor 7 calculates the probability that each imaging part is included in the input image 96. Then, the processor 7 infers the imaging part included in the input image 96 based on the probability calculated for each imaging part.

[0253] Here, it is assumed that the probability of the kidney exceeds the threshold value. Therefore, the processor 7 infers that the imaging region included in the input image 96 is the kidney. After inferring the imaging region, the process proceeds to step ST30.

[0254] In step ST30, the processor 7 determines whether or not to change the imaging conditions based on the inferred imaging region. Step ST30 will be specifically described below.

[0255] First, in step ST31, the processor 7 judges whether the currently set imaging conditions are imaging conditions corresponding to the imaging region inferred in step ST12. At time t7, the set imaging conditions are imaging conditions for the liver (N). On the other hand, the imaging region inferred in step ST12 is the kidney. Therefore, at time t7, the set imaging conditions (imaging conditions for the liver (N)) are not imaging conditions corresponding to the imaging region (kidney) inferred in step ST12, so the process proceeds to step ST33.

[0256] In step ST33, as shown in Fig. 22, the processor 7 identifies a category corresponding to the inferred imaging region from among a plurality of categories 1 to N registered in the database. Here, since the inferred imaging region is "kidney", the processor 7 identifies category 4 including a parameter set for kidney from among the plurality of categories 1 to N.

[0257] Next, the processor 7 identifies a category corresponding to the currently set imaging condition from among a plurality of categories 1 to N registered in the database. Here, since the currently set imaging condition is an imaging condition for the liver (N), the processor 7 identifies category 4 including a parameter set for the liver (N) from among a plurality of categories 1 to n. After identifying category 4, the process proceeds to step ST34.

[0258] In step ST34, the processor 7 determines whether the two categories identified in step ST33 match. If the categories match, the process proceeds to step ST32, whereas if the categories do not match, the process proceeds to step ST35. Here, the identified categories match (category 4), so the process proceeds to step ST35.

[0259] In step ST35, the processor 7 determines to change the shooting conditions. That is, when it is determined that the categories do not match, the processor 7 determines to change the shooting conditions. The reason for this will be described below.

[0260] Category 4 includes a liver (N) parameter set D1, a liver (L) parameter set D2, a kidney parameter set D3, and an intestine parameter set D4. The liver, kidney, and intestine, which are the imaging targets of these parameter sets, are anatomically different organs. However, the values ​​of the parameters used in imaging the liver, kidney, and intestine are often the same or close to each other. Therefore, it is considered that the image quality of the ultrasound image is not significantly affected even if an ultrasound image of the kidney is acquired without changing the currently set imaging condition Y1 for the liver (N) to the inferred imaging condition for the kidney, and the imaging condition for the liver (N) is kept as it is. Therefore, it is determined not to change the imaging condition in step ST32, and the process 44 is terminated.

[0261] Therefore, the user 51 continues the examination of the kidney of the subject S43 under the imaging conditions for the liver (N) even after time t7. Meanwhile, the processor 7 periodically executes the process 44 even after time t7. Then, when the acquisition of the necessary images is completed, the examination of the kidney of the subject S43 ends (time t8).

[0262] In the fourth embodiment, the category 4 of the database includes the liver (N) parameter set D1, the liver (L) parameter set D2, the kidney parameter set D3, and the intestine parameter set D4. The liver (N) and the liver (L) are common to the liver as an organ. Although the liver, the kidney, and the intestine are different organs, there is not much difference in the imaging conditions. Therefore, for the liver (N), the liver (L), the kidney, and the intestine, an ultrasonic image of a certain quality can be obtained by using any of the parameter sets among the liver (N) parameter set D1, the liver (L) parameter set D2, the kidney parameter set D3, and the intestine parameter set D4. Therefore, in the fourth embodiment, the liver (N) parameter set D1, the liver (L) parameter set D2, the kidney parameter set D3, and the intestine parameter set D4 are classified into the same category 4. By classifying the parameter sets D1 to D4 into the same category 4, the imaging conditions are not changed between the liver (N), the liver (L), the kidney, and the intestine, so that the frequency of the imaging conditions being automatically changed at a timing unintended by the user 51 can be reduced. Furthermore, since the liver (N) parameter set D1, liver (L) parameter set D2, kidney parameter set D3, and intestine parameter set D4 have the same or similar parameter values, ultrasound images of sufficient quality can be obtained even if the shooting conditions are not changed between the liver (N), liver (L), kidneys, and intestines. [Explanation of symbols]

[0263] 1. Ultrasound diagnostic equipment 2 Ultrasound probe 3. Transmit Beamformer 4 Transmitter 5 Receiver 6 Receive beamformer 7 Processors 8 Display section 9. Memory 10 User Interface 18 Display Monitor 28 Touch Panel 30 Neural Networks 31, 32, 33 Pre-trained models 41, 42, 43, 44 Process 51 users 61, 63, 65, 71, 73, 75, 81, 83, 85, 91, 93, 95, Ultrasound images 62, 64, 66, 72, 76, 82, 84, 86, 92, 94, 96 Input image

Claims

1. An ultrasonic probe; a processor in communication with the ultrasound probe; An ultrasound diagnostic device comprising: The processor, Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run The processor, When it is determined that the ultrasound image includes a plurality of imaging regions, the ultrasound diagnostic apparatus determines not to change the imaging conditions.

2. The processor, generating an input image to be input to a trained model based on the ultrasound image; Inputting the input image into the trained model to infer an imaging site included in the input image; determining whether the number of imaging regions inferred for the input image is one or more than one; When it is determined that the number of imaging regions inferred for the input image is two or more, it is determined that the imaging conditions are not to be changed. The ultrasonic diagnostic apparatus according to claim 1 ,

3. The processor, When it is determined that the number of imaging regions inferred for the input image is one, determining whether or not to change the imaging conditions. The ultrasonic diagnostic apparatus according to claim 2 ,

4. An ultrasonic probe; a processor in communication with the ultrasound probe and a storage device; 1. An ultrasound system comprising: The processor, Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run The storage device stores a database including a plurality of categories, each of the plurality of categories including a plurality of parameter sets; The processor, identifying a first category corresponding to the determined imaging region from among the plurality of categories; identifying a second category corresponding to the set photographing condition from among the plurality of categories; determining not to change the photographing condition if the first category and the second category match; An ultrasound diagnostic device that performs the above steps.

5. The processor, generating an input image to be input to a trained model based on the ultrasound image; Inputting the input image into the trained model to infer an imaging site included in the input image; The ultrasonic diagnostic apparatus according to claim 4 ,

6. The processor, Determining whether the set imaging condition is an imaging condition corresponding to the inferred imaging region; and determining not to change the imaging condition when the set imaging condition corresponds to the inferred imaging region; The ultrasonic diagnostic apparatus according to claim 5 ,

7. The processor, The ultrasonic diagnostic apparatus according to claim 6 , wherein the first category and the second category are identified when the set imaging condition is not an imaging condition corresponding to an inferred imaging region.

8. The processor, The ultrasound diagnostic apparatus according to claim 7 , wherein when the first category and the second category do not match, it is determined that the imaging condition is to be changed.

9. The ultrasonic diagnostic apparatus according to claim 4 , wherein the parameter set includes at least a frequency or a depth.

10. 2. The ultrasound diagnostic apparatus according to claim 1, wherein the plurality of parameter sets included in each of the plurality of categories are parameter sets for a plurality of regions defined as having similar body compositions of the subject, or parameter sets set according to different body compositions in one region.

11. 1. A recording medium having stored thereon instructions executable by a processor in communication with an ultrasound probe, the instructions causing the processor to: Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run the command, The instructions further include for the processor to: a storage medium for causing a decision not to change the imaging conditions when it is determined that the ultrasound image includes a plurality of imaging regions;

12. 1. A recording medium having stored thereon instructions executable by a processor in communication with an ultrasound probe, the instructions causing the processor to: Setting imaging conditions for acquiring an ultrasound image of the subject; causing the ultrasonic probe to transmit an ultrasonic beam in accordance with the imaging conditions, causing the ultrasonic probe to receive echoes from the subject, and generating an ultrasonic image of the subject based on the echoes received by the ultrasonic probe; Identifying an imaging region included in the ultrasound image; determining whether to change the imaging conditions based on the determined imaging region; Run the command, The determination of whether to change the imaging condition is identifying a first category corresponding to the determined imaging region from among the plurality of categories; identifying a second category corresponding to the set photographing condition from among the plurality of categories; determining not to change the photographing condition if the first category and the second category match; A storage medium including

Images